At some point, if we want this integration to move beyond proof-of-concept, we'll need to install and maintain our own SonarQube instance.
Our (Code-Health-Metrics) current thinking is:
- continue to use sonarcloud.io for per-patch analysis and on postmerge
- also on postmerge, analyze patches with our self-hosted SonarQube instance
That way, we will maintain control of our historical data in our self-hosted instance. We can't use the self-hosted instance for per-patch analysis without paying a lot of money for a license.